Page 1 of 2 12 LastLast
Results 1 to 10 of 19

Thread: Davis values

  1. #1
    Join Date
    Sep 2006
    Location
    .
    Posts
    2,491

    Default Davis values

    Having noted how the last few releases from Trainsimulations (Mullan Pass, Rock Island) have reverted to MSTS physics instead of using Davis values (ORTS Selgman, Ruel, Milw) I wanted to see what the actual difference between the two sets of figures were demonstrating. A little bit of testing on flat level track shows how the Davis numbers were making the cars a bit too 'sticky' when stationary, requiring much more power to simply get them rolling.

    I created two 90 car trains of 286k covered hoppers (about 12,500 US tons) and attached a single SD40-2. The brown BNSF cars are from Seligman 2 route and have Davis values. The identical gray CEFX cars are from Mullan route and have MSTS Friction values. The only editing done to wag files was to reduce the mass for the seligman car to match that of the mullan car.

    With a solid train of Davis values, I could not get the train moving with the single SD40-2. Only after I reduced the train length to 1/2 the original size (down to 45 cars) while the engine sat there in Run 8 did the train finally begin to crawl forward. Once it did get the train rolling, it got out of the red zone faster on account of the shorter consist.

    https://www.youtube.com/watch?v=7Jnz1wHf1dI

    Compare to the CEFX consist that utilizes MSTS friction values. The lone SD40-2 was capable of getting the entire 90 car train moving.

    https://www.youtube.com/watch?v=TTKNPSQUuiA

    I can understand how the real world train weights, the 'doubles', and helper assignments found in mountainous western US railroading would have been very hard to replicate successfully when using the stickier Davis values. Even in eastern railroading, how many years did Conrail dispatch 110 car coal trains between Altoona and Harrisburg with only 2 older SD40 types for power? I'm sure at some point engine failure would have resulted in only one SD40 bringing the train home to the end terminal.

    Sometimes we get a little too involved in the math that we really don't see how our end result fails at actually mirroring real world conditions.

  2. #2
    Join Date
    Feb 2013
    Location
    known universe
    Posts
    2,043

    Default

    Interesting, but will withhold comment until I see all the Davis information, the values, and how they were calculated. Care to provide them?
    Providing consist info would also be helpful, to duplicate your results. You may, or may not have a point. Regards, Gerry
    Cheers, Gerry
    It's my railroad and I'll do what I want! Historically accurate attitude of US Railroad Barons.
    Forever, ridin' drag in railroad knowledge.


  3. #3
    Join Date
    Jul 2004
    Location
    Wareham, Dorset, U.K.
    Posts
    2,901

    Default

    Quote Originally Posted by geepster775 View Post
    Sometimes we get a little too involved in the math that we really don't see how our end result fails at actually mirroring real world conditions.
    Well said Geepster! WAY TOO often do the theorists force their opinions on technical hobbies, apparently blind to the realities around them. Those of my generation when HiFi was all the rage will remember scornful reviews of equipment for failing to measure up on instruments to the "so called" best buys. Quite often the differences could ONLY be observed on instruments because they were way beyond even the very best human hearing

    Differences between the Sim's physics models aside I've often looked at the many books in my Railroad library and looked at pictures of loco's back in the 70's where two loco's were obviously managing trains so long they disappeared from the camera's viewpoint. I always thought "I bet the Sim couldn't replicate THAT".

    Maybe it's been the obsession with building routes with extreme gradients? Nearly ALL payware routes have been guilty of this from Kicking Horse Pass onwards. There's no doubt that grades and twisting valley runs are a challenging drive but in reality what % mileage of North American track is through these?

    I've just downloaded the latest version of the SAL route (mentioned recently in the forums) and whizzing along speeded up in OR it to see how good the scenery was and what the route went through I was immediately struck by how much of the Rail in almost flat Florida was actually straight as a die.

    I think you may have hit a few nerves with this post ...but I'm no physics authority so I'll just sit back and watch them justify why the "advanced" physics are simply not realistic.

    Don't even get me started on braking because that's one aspect that is WAY OUT in nearly all physics models.
    Geoff
    Dorset - near The Swanage Railway.
    UK

  4. #4
    Join Date
    May 2005
    Location
    Stalinist,broke United Kingdom (Hey buddy,can you spare �1.3 trillion?))
    Posts
    2,138

    Default

    I can comment only on MSTS,but most models brake force values are far too low.

  5. #5
    Join Date
    Feb 2013
    Location
    known universe
    Posts
    2,043

    Default

    Quote Originally Posted by lateagain View Post
    Well said Geepster!...and
    I think you may have hit a few nerves with this post ...but I'm no physics authority so I'll just sit back and watch them justify why the "advanced" physics are simply not realistic.
    Yep, I always read geepster posts because I respect the member's experience and opinion. I am only asking for work product so I can set up and perform exactly the same tests...this is about replicating ( and therefore verifying ) the results he got. My initial reaction is that the interpretation geepster provided is correct...I just want to investigate for myself.

    Contact me by PM Geepster, thanks for the observations, somewhat aligns with some things I've seen. -- maybe there is something else going on.

    Geoff, I also enjoy your efforts to get everyone's "socks in a knot"...you gotta get your entertainment where you can find it.
    Cheers, Gerry
    It's my railroad and I'll do what I want! Historically accurate attitude of US Railroad Barons.
    Forever, ridin' drag in railroad knowledge.


  6. #6
    Join Date
    Jul 2004
    Location
    Wareham, Dorset, U.K.
    Posts
    2,901

    Default

    Quote Originally Posted by R. Steele View Post
    Geoff, I also enjoy your efforts to get everyone's "socks in a knot"...you gotta get your entertainment where you can find it.
    I never have to put ANY effort in Gerry.... many of life's "experts" find it hard to be open minded ...and that's in any walk of life.

    ...and the braking was not a wind up. Most braking is totally unrealistic on far too many models.
    Geoff
    Dorset - near The Swanage Railway.
    UK

  7. #7
    Join Date
    Feb 2013
    Location
    known universe
    Posts
    2,043

    Default

    Quote Originally Posted by lateagain View Post
    ...and the braking was not a wind up. Most braking is totally unrealistic on far too many models.
    Oh, quite agree there, have you had the opportunity to get a copy of Joe Realmuto's pdf on "Guideline For Setting Open Rails Braking Parameters (Including Blended Braking)"?
    Cheers, Gerry
    It's my railroad and I'll do what I want! Historically accurate attitude of US Railroad Barons.
    Forever, ridin' drag in railroad knowledge.


  8. #8
    Join Date
    Nov 1999
    Location
    Doreen, Victoria, Australia.
    Posts
    6,762

    Default

    I have a vague memory that the Davis figures does not get the 0-5kmh friction correct, I am sure it has been discussed before.
    When I find it, I will post here.

    found it


    My loaded 45 ton coal wagon shows a standing resistance of 971N using the standard friction of 492.2N/m/s in the wag file.
    When using the ORTSDavis friction parameters for ORTSDavis_A ( 492.23 ). The standing resistance is 2073N.
    I cannot seem to adjust the standing resistance using the ORTS davis parameters.
    Does anyone know which of the two standing resistance above, is correct for a 2-axle 45t coal wagon ?

    Thanks


    Posted 08 July 2018 - 06:12 AM

    The problem is neither Kuju's friction formula or any of the Davis formulas compute resistance under 5mph -- the values shoot up as speed drops under 5mph. 492.2N/m/s in friction looks very light to me. What were all the values of friction beforehand? Keep in mind that just because somebody put values into Friction() before sending the file to you doesn't mean they're anywhere outside of complete fantasy. For many years I've used a spreadsheet to graph the difference between friction and the Davis curve and checked many downloaded .wags... some of the friction stuff has been outright bizarre.

    My recollection of both formulas is the KUJU resistance values get real hinky when you move the Vel2 value of friction above 5 and/or the value of Exp1 above 2.65... you get these big sawtooth spikes in the calculated curve in the first 5-6 mph. Drop either con1 or Exp1 too low and the bottom drops out such that your grandmother could move the whole train on her own.

    So w/o seeing all of your data I'm inclined to trust fcalc2 more than Kuju. Perhaps you should revisit it just on the chance a keystroke was in error. Or give me the old friction values and I'll tell you if they were reasonable.
    Dave Nelson


    Posted 08 July 2018 - 11:15 AM

    The resistance friction figures are taken from the Fcalc2 program! For a 2 axle 45 ton hopper wagon, Fcalc2 and with the Davis A formula both give 492.

    Without the Davis parameters, OR reports the standing friction as 971N. With the Davis parameter, the standing resistance is 2073N.

    All my trainset use Fcalc2 with the required parameters. Only the Davis C and last entry on the kuju friction parameters are adjusted for wind resistance at high speed.



    There are other in depth discussions about friction over at elvas tower.
    Last edited by derekmorton; 09-20-2020 at 12:49 AM.
    Cheers
    Derek

  9. #9

    Default

    Quote Originally Posted by geepster775 View Post
    Having noted how the last few releases from Trainsimulations (Mullan Pass, Rock Island) have reverted to MSTS physics instead of using Davis values.
    In testing with the creator of Mullan Pass, I believe that we were able to achieve realistic operation with the Davis values (including the starting resistance) against real life tonnage ratings.

    This was after some changes were made to how power was calculated in the diesel model for OR. These changes are now in the current unstable version. (If you want to read some of the background, as well as the testing that was done supporting it, see this thread - Diesel Locomotive Performance ).

    My understanding of the main reason that Mullan Pass opted for the MSTS values was because they wanted to use the latest STABLE version of OR, and not take a risk that users were using an earlier version of OR without the relevant changes in place. Hence it gave them a "stable" platform to work with, and reduced any potential user problems.

  10. #10

    Default

    Please do tell -
    Some of us have just spent months creating OR folders with OR eng and wag files that use Davis values, with the eng files also referencing the set of inc files created by Jerry Storey. Are we going to have to revise all that?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •